inra.ijpb.watershed

Class MarkerControlledWatershedTransform2D

java.lang.Object

inra.ijpb.watershed.WatershedTransform2D

inra.ijpb.watershed.MarkerControlledWatershedTransform2D

public class MarkerControlledWatershedTransform2D
extends WatershedTransform2D

Marker-controlled version of the watershed transform in 2D. References: [1] Fernand Meyer and Serge Beucher. "Morphological segmentation." Journal of visual communication and image representation 1.1 (1990): 21-46. [2] Peer Neubert and Peter Protzel. "Compact Watershed and Preemptive SLIC: On improving trade-offs of superpixel segmentation algorithms." 22nd international conference on pattern recognition. IEEE, 2014.

Author:

Ignacio Arganda-Carreras

Field Summary

Fields inherited from class inra.ijpb.watershed.WatershedTransform2D

verbose

Constructor Summary

Constructors

Constructor and Description
MarkerControlledWatershedTransform2D(ij.process.ImageProcessor input, ij.process.ImageProcessor marker, ij.process.ImageProcessor mask) Initialize a marker-controlled watershed transform
MarkerControlledWatershedTransform2D(ij.process.ImageProcessor input, ij.process.ImageProcessor marker, ij.process.ImageProcessor mask, double compactness) Initialize a marker-controlled watershed transform.
MarkerControlledWatershedTransform2D(ij.process.ImageProcessor input, ij.process.ImageProcessor marker, ij.process.ImageProcessor mask, int connectivity) Initialize a marker-controlled watershed transform
MarkerControlledWatershedTransform2D(ij.process.ImageProcessor input, ij.process.ImageProcessor marker, ij.process.ImageProcessor mask, int connectivity, double compactness) Initialize a marker-controlled watershed transform.

Method Summary

Modifier and Type	Method and Description
ij.process.ImageProcessor	applyWithPriorityQueue() Apply watershed transform on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage.
ij.process.ImageProcessor	applyWithPriorityQueueAndDams() Apply watershed transform on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage (optionally).
ij.process.ImageProcessor	applyWithSortedList() Deprecated. The algorithm with a sorted list does not visit the pixels based on their h value and proximity to markers so it is not a true watershed method.
ij.process.ImageProcessor	applyWithSortedListAndDams() Deprecated. The algorithm with a sorted list does not visit the pixels based on their h value and proximity to markers so it is not a true watershed method.
LinkedList<PixelRecord>	extractPixelValues(ij.process.ImageProcessor inputImage, ij.process.ImageProcessor markerImage, int[][] tabLabels) Extract pixel values from input and labeled marker images.
PriorityQueue<PixelRecord>	extractPixelValuesPriorityQueue(ij.process.ImageProcessor inputImage, ij.process.ImageProcessor seedImage, int[][] tabLabels) Extract pixel values from input and seed images
ij.ImagePlus	getAnimationPriorityQueue() Get animation of each h-step of the watershed transform on the inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage.
ij.ImagePlus	getAnimationPriorityQueueAndDams() Get an animation of the different steps of the watershed transform applied on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage (optionally).
ij.ImagePlus	getAnimationSortedList() Deprecated. The algorithm with a sorted list does not visit the pixels based on their h value and proximity to markers so it is not a true watershed method.
ij.ImagePlus	getAnimationSortedListAndDams() Deprecated. The algorithm with a sorted list does not visit the pixels based on their h value and proximity to markers so it is not a true watershed method.

Methods inherited from class inra.ijpb.watershed.WatershedTransform2D

apply, apply, extractPixelValues, getAnimation, getConnectivity, setConnectivity, setVerbose

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

MarkerControlledWatershedTransform2D

public MarkerControlledWatershedTransform2D(ij.process.ImageProcessor input,
                                            ij.process.ImageProcessor marker,
                                            ij.process.ImageProcessor mask)

Initialize a marker-controlled watershed transform

Parameters:

input - grayscale image (usually a gradient image)

marker - image containing the labeled markers to start the watershed

mask - binary mask to restrict the region of interest (null to use whole input image)

MarkerControlledWatershedTransform2D

public MarkerControlledWatershedTransform2D(ij.process.ImageProcessor input,
                                            ij.process.ImageProcessor marker,
                                            ij.process.ImageProcessor mask,
                                            int connectivity)

Initialize a marker-controlled watershed transform

Parameters:

input - grayscale image (usually a gradient image)

marker - image containing the labeled markers to start the watershed

mask - binary mask to restrict the region of interest (null to use whole input image)

connectivity - pixel connectivity (4 or 8)

MarkerControlledWatershedTransform2D

public MarkerControlledWatershedTransform2D(ij.process.ImageProcessor input,
                                            ij.process.ImageProcessor marker,
                                            ij.process.ImageProcessor mask,
                                            double compactness)

Initialize a marker-controlled watershed transform. If the compactness constraint is larger than 0, the Compact Watershed algorithm will be executed (Peer Neubert and Peter Protzel. "Compact Watershed and Preemptive SLIC: On improving trade-offs of superpixel segmentation algorithms." 22nd international conference on pattern recognition. IEEE, 2014).

Parameters:

input - grayscale image (usually a gradient image)

marker - image containing the labeled markers to start the watershed

mask - binary mask to restrict the region of interest (null to use whole input image)

compactness - compactness constrain parameter (values larger than 0 imply using compact watershed)

MarkerControlledWatershedTransform2D

public MarkerControlledWatershedTransform2D(ij.process.ImageProcessor input,
                                            ij.process.ImageProcessor marker,
                                            ij.process.ImageProcessor mask,
                                            int connectivity,
                                            double compactness)

Initialize a marker-controlled watershed transform. If the compactness constraint is larger than 0, the Compact Watershed algorithm will be executed (Peer Neubert and Peter Protzel. "Compact Watershed and Preemptive SLIC: On improving trade-offs of superpixel segmentation algorithms." 22nd international conference on pattern recognition. IEEE, 2014).

Parameters:

input - grayscale image (usually a gradient image)

marker - image containing the labeled markers to start the watershed

mask - binary mask to restrict the region of interest (null to use whole input image)

connectivity - pixel connectivity (4 or 8)

compactness - compactness constrain parameter (values larger than 0 imply using compact watershed)

Method Detail

applyWithSortedList

@Deprecated
public ij.process.ImageProcessor applyWithSortedList()

Deprecated. The algorithm with a sorted list does not visit the pixels based on their h value and proximity to markers so it is not a true watershed method.

Apply watershed transform on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage. This implementation visits all pixels by ascending gray value.

Returns:

watershed domains image (no dams)

getAnimationSortedList

@Deprecated
public ij.ImagePlus getAnimationSortedList()

Deprecated. The algorithm with a sorted list does not visit the pixels based on their h value and proximity to markers so it is not a true watershed method.

Get animation of the different steps in the watershed transform on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage. This implementation visits all pixels by ascending gray value.

Returns:

animation of watershed domains image (no dams)

applyWithSortedListAndDams

@Deprecated
public ij.process.ImageProcessor applyWithSortedListAndDams()

Deprecated. The algorithm with a sorted list does not visit the pixels based on their h value and proximity to markers so it is not a true watershed method.

Apply watershed transform on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage. This implementation visits all pixels by ascending gray value.

Returns:

watershed domains image (with dams)

getAnimationSortedListAndDams

@Deprecated
public ij.ImagePlus getAnimationSortedListAndDams()

Deprecated. The algorithm with a sorted list does not visit the pixels based on their h value and proximity to markers so it is not a true watershed method.

Get animation of the different steps in the watershed transform applied to the inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage. This implementation visits all pixels by ascending gray value.

Returns:

image stack with the steps of the watershed transform (with dams)

applyWithPriorityQueue

public ij.process.ImageProcessor applyWithPriorityQueue()

Apply watershed transform on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage. This implementation visits first the pixels on the surroundings of the labeled markers.

Returns:

watershed domains image (no dams)

getAnimationPriorityQueue

public ij.ImagePlus getAnimationPriorityQueue()

Get animation of each h-step of the watershed transform on the inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage. This implementation visits first the pixels on the surroundings of the labeled markers.

Returns:

animation of the watershed domains image (no dams)

getAnimationPriorityQueueAndDams

public ij.ImagePlus getAnimationPriorityQueueAndDams()

Get an animation of the different steps of the watershed transform applied on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage (optionally). This implementation uses a priority queue to visit first the pixels on the surroundings of the labeled markers (Meyer's flooding algorithm). Meyer's flooding algorithm: Label the regional minima with different colors Repeat | Select a pixel p, not colored, not watershed, | adjacent to some colored pixels, | and having the lowest possible gray level | If p is adjacent to exactly one color then | label p with this color | If p is adjacent to more than one color then | label p as watershed Until no such pixel exists More information at - Serge Beucher's site: http://cmm.ensmp.fr/~beucher/wtshed.html - G. Bertrand's Topological Watershed site: http://www.esiee.fr/~info/tw/index.html

Returns:

animation stack with watershed domains by each h level (with dams)

applyWithPriorityQueueAndDams

public ij.process.ImageProcessor applyWithPriorityQueueAndDams()

Apply watershed transform on inputImage, using the labeled markers from markerImage and restricted to the white areas of maskImage (optionally). This implementation uses a priority queue to visit first the pixels on the surroundings of the labeled markers (Meyer's flooding algorithm). Meyer's flooding algorithm: Label the regional minima with different colors Repeat | Select a pixel p, not colored, not watershed, | adjacent to some colored pixels, | and having the lowest possible gray level | If p is adjacent to exactly one color then | label p with this color | If p is adjacent to more than one color then | label p as watershed Until no such pixel exists More information at - Serge Beucher's site: http://cmm.ensmp.fr/~beucher/wtshed.html - G. Bertrand's Topological Watershed site: http://www.esiee.fr/~info/tw/index.html

Returns:

watershed domains image (with dams)

extractPixelValuesPriorityQueue

public PriorityQueue<PixelRecord> extractPixelValuesPriorityQueue(ij.process.ImageProcessor inputImage,
                                                                  ij.process.ImageProcessor seedImage,
                                                                  int[][] tabLabels)

Extract pixel values from input and seed images

Parameters:

inputImage - input image (2D)

seedImage - seed image (2D)

tabLabels - output label array

Returns:

priority queue of pixels neighboring the seeds

extractPixelValues

public LinkedList<PixelRecord> extractPixelValues(ij.process.ImageProcessor inputImage,
                                                  ij.process.ImageProcessor markerImage,
                                                  int[][] tabLabels)

Extract pixel values from input and labeled marker images. The input grayscale values will be return in a list of PixelRecord and the markers will be stored in tabLabels.

Parameters:

inputImage - input grayscale image (usually a gradient image)

markerImage - labeled marker image

tabLabels - output label array

Returns:

list of input pixel values